Contribution of genetic and epigenetic mechanisms to Wnt

1Q2

2

3Q1

4

5

67891011

12

13141516171819202122232425

47

48

Gene xxx (2013) xxx–xxx

GENE-39080; No. of pages: 8; 4C:

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r .com/ locate /gene

Contribution of genetic and epigenetic mechanisms to Wnt pathwayactivity in prevalent skeletal disorders

RO

OFCarmen García-Ibarbia a, Jesús Delgado-Calle a, Iñigo Casafont b, Javier Velasco a, Jana Arozamena a,

María I. Pérez-Núñez c, María A. Alonso c, María T. Berciano b, Fernando Ortiz a, José L. Pérez-Castrillón d,Agustín F. Fernández e, Mario F. Fraga e, María T. Zarrabeitia f, José A. Riancho a,⁎a Department of Internal Medicine, Hospital U.M. Valdecilla-IFIMAV, University of Cantabria, RETICEF, Santander, Spainb Department of Anatomy and Cell Biology, University of Cantabria, IFIMAV, Santander, Spainc Department of Orthopaedic Surgery and Traumatology, Hospital U.M. Valdecilla, Santander, Spaind Department of Internal Medicine, Hospital U, Rio-Hortega, University of Valladolid, Valladolid, Spaine Cancer Epigenetics Laboratory, Instituto Universitario de Oncología del Principado de Asturias (IUOPA), HUCA, University of Oviedo, Oviedo, Spainf Unit of Legal Medicine, University of Cantabria, IFIMAV, Santander, Spain

UNC

Abbreviations: AzadC, 5-aza-2-deoxy-azacytidine; CACCAMK2G, calcium/calmodulin-dependent protein kinasenase 1, alpha 1; CSNK1A1L, casein kinase 1, alpha 1-like; CCt, threshold cycle; CTBP1, C-terminal binding protein 1; FFOS-like antigen1; FRZB, frizzled-related protein; FZD10, fcogen synthase kinase 3 beta; GWAS, genome-wide asWeinberg equilibrium; LRP5, lipoprotein receptor related pC, beta 3 (phosphatidylinositol-specific); PPP2R1A, proteiregulatory subunit A, alpha isoform; RHOA, ras homologsecreted frizzled-related protein 1; SFRP4, secreted friztransducin (beta)-like 1X-linked; TBP, TATA box bindintype MMTV integration site family, member 10A; WNT16tion site family, member 16;WNT4, wingless-typeMMTV4; WNT8A, wingless-type MMTV integration site family, m⁎ Corresponding author at: Department of Internal Me

University of Cantabria, IFIMAV, Santander 39008, Sfax: +34 942201695.

E-mail address: [email protected] (J.A. Riancho).

0378-1119/$ – see front matter © 2013 Published by Elsehttp://dx.doi.org/10.1016/j.gene.2013.09.080

Please cite this article as: García-Ibarbia, C.skeletal disorders, Gene (2013), http://dx.do

Pa b s t r a c t
a r t i c l e i n f o
26

27

28

29

30

31

32

33

34

Article history:Accepted 23 September 2013Available online xxxx

Keywords:Wntβ-CateninDNA methylationFracturesBone diseases

35

36

37

38

39

40

41

42

43

44

RRECTEDWe reported previously that the expression of Wnt-related genes is lower in osteoporotic hip fractures than in

osteoarthritis. We aimed to confirm those results by analyzing β-catenin levels and explored potential geneticand epigenetic mechanisms involved.β-Catenin gene expression and nuclear levelswere analyzed by real time PCR and confocal immunofluorescence.Increased nuclear β-catenin was found in osteoblasts isolated from patients with osteoarthritis (99 ± 4units vs. 76 ± 12, p = 0.01, n = 10), without differences in gene transcription, which is consistent witha post-translational down-regulation of β-catenin and decreased Wnt pathway activity.Twenty four single nucleotide polymorphisms (SNPs) of genes showing differential expression between fracturesand osteoarthritis (WNT4, WNT10A, WNT16 and SFRP1) were analyzed in DNA isolated from blood of 853 pa-tients. The genotypic frequencies were similar in both groups of patients, with no significant differences.Methylation of Wnt pathway genes was analyzed in bone tissue samples (15 with fractures and 15 with osteo-arthritis) by interrogating a CpG-based methylation array. Six genes showed significant methylation differencesbetween both groups of patients: FZD10, TBL1X, CSNK1E, WNT8A, CSNK1A1L and SFRP4. The DNA demethylatingagent 5-deoxycytidine up-regulated 8 genes, including FZD10, in an osteoblast-like cell line, whereas it down-regulated other 16 genes.In conclusion,Wnt activity is reduced in patientswith hip fractures, in comparisonwith thosewith osteoarthritis.It does not appear to be related to differences in the allele frequencies of the Wnt genes studied. On the otherhand, methylation differences between both groups could contribute to explain the differences in Wnt activity.

© 2013 Published by Elsevier B.V.

4546
O
49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

YBP, calcyclin binding protein;II gamma; CSNK1A1, casein ki-SNK1E, casein kinase 1, epsilon;DR, false discovery rate; FOSL1,rizzled homolog 10; GSK3B, gly-sociation study; HWE, Hardy–rotein 5; PLCB3, phospholipasen phosphatase 2 (formerly 2A),gene family, member A; SFRP1,zled-related protein 4; TBL1X,g protein; WNT10A, wingless-, wingless-type MMTV integra-integration site family, memberember 8A.

dicine, Hospital U.M. Valdecilla,pain. Tel.: +34 942201990;

vier B.V.

, et al., Contribution of geneti.org/10.1016/j.gene.2013.09.

1. Introduction

TheWnt pathway has emerged as an important regulator of skeletalhomeostasis. Binding of Wnt ligands to their receptors triggers the acti-vation of a complex signaling pathway. Multiple intracellular mediatorsare involved, but the best known cascade ofWnt signals constitutes theso-called canonical pathway, which involves the post-transcriptionalregulation of β-catenin levels. Wnt ligands induce the disassemblyof the GSK3 complex that phosphorylates β-catenin. Since non-phosphorylated β-catenin is less prone to proteasome degradation,this results in increased β-catenin levels and translocation into thenucleus, where it modulates the transcription of target genes (Gauret al., 2005; Williams and Insogna, 2009), with the collaboration ofseveral co-factors, including members of the T-cell factor/lymphoidenhancing factor family (Gordon and Nusse, 2006; MacDonald et al.,2009). There are 19 different Wnt ligands, some of which, such as

ic and epigenetic mechanisms to Wnt pathway activity in prevalent080

http://dx.doi.org/10.1016/j.gene.2013.09.080

mailto:[email protected]


http://www.sciencedirect.com/science/journal/03781119


T

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

2 C. García-Ibarbia et al. / Gene xxx (2013) xxx–xxx

UNCO

RREC

Wnt3a, preferentially activate the canonical pathway, whereas others,such as Wnt4 and Wnt5a, are usually regarded to transmit signalsthrough non-canonical pathways. However, there is no a clear differ-ence between ligands, each one being able to activate preferentiallythe canonical or non-canonical pathways depending on the target celland other context-dependent factors (van Amerongen and Nusse,2009). Furthermore, cross-talks take place between different pathways.Both the canonical and the non-canonical pathways appear to be in-volved in the regulation of bone homeostasis (Chang et al., 2007; Gauret al., 2005; Piters et al., 2008). Wnt inhibitors include members of thesecreted frizzled related protein family, which are structurally relatedto the Wnt membrane-bound receptors frizzled and bind directly toWnt ligands, thereby altering their ability to interact with the Wnt re-ceptor complex at cell membranes (Kawano and Kypta, 2003). Onemember of this family, sFRP1, has been shown to modulate the activityof cells of the osteoblastic lineage (Bodine et al., 2005; Yao et al., 2010).In line with this, knock-out mice with deletion of the SFRP1 or othergenes encoding secreted frizzled proteins show increased bone mass(Lodewyckx and Lories, 2009).

An association between Wnt-related genes, particularly FRZB, withosteoarthritis was reported in several studies (Loughlin et al., 2004;Valdes et al., 2007), but it could not be confirmed in recent GWAS(Arcogen Consortium, 2012; Panoutsopoulou et al., 2011). On theother hand, variations in genes related to the Wnt pathway, such asthe Wnt co-receptors LRP5 and the Wnt inhibitor sclerostin, havebeen related to osteoporosis in genetic association studies (Ralston,2010; Riancho et al., 2011; Richards et al., 2009; Styrkarsdottir et al.,2009; Valero et al., 2011). Specifically, Estrada et al. found an associationbetween some polymorphisms of genes encoding Wnt ligands, such asWNT16, and osteoporosis in a large multinational GWAS (Estradaet al., 2012).

Epigenetic mechanisms, and specifically the methylation of CpGsites in gene promoters, are known to play an important role in gene ex-pression regulation during development and in adult organisms (Fragaand Esteller, 2007). The role of DNA methylation in bone homeostasishas not been extensively studied yet, but several lines of evidencepoint it as a critical regulator of the differentiation of bone cells(Delgado-Calle et al., 2012a,c).

Osteoporosis and osteoarthritis are prevalent skeletal disorders.Whereas bone mass is decreased in osteoporosis, several epidemio-logical studies suggested that patients with osteoarthritis may shownot only periarticular bone formation, but also a generalized trendfor higher bone mass (Arokoski et al., 2002; Chaganti et al., 2010;Dequeker et al., 2003).We have previously reported that, in compar-ison with patients with osteoarthritis, the expression of a number ofgenes in theWnt pathway is reduced in bone samples and osteoblastcultures from patients with osteoporotic hip fractures. In line withthis, experiments with a reporter vector suggested higherWnt activ-ity is osteoarthritis, suggesting that differences in Wnt activity maybe involved in the opposite changes in bone mass typical of thesedisorders (Velasco et al., 2010). In the present study we comparednuclear β-catenin levels in primary osteoblasts from patients withhip fractures and hip osteoarthritis, determined β-catenin geneexpression, and explored if the differences in Wnt activity were re-lated to genotypic or epigenetic differences between both groupsof patients.

178

179

180

181

182

183

184

185

2. Materials and methods

2.1. Patients

The study subjects included patients with osteoporotic hip fracturesor with severe hip osteoarthritis requiring replacement surgery. Sam-ples from different patients were used for the various experiments. Pa-tients with secondary osteoarthritis or secondary osteoporosis, those

Please cite this article as: García-Ibarbia, C., et al., Contribution of genetskeletal disorders, Gene (2013), http://dx.doi.org/10.1016/j.gene.2013.09.

ED P

RO

OF

taking drugs known to affect bone metabolism, as well as those havingfractures related to high-energy trauma, were excluded.

Patients gave informed consent. The study was approved by theinstitutional review board (Comité de Etica en Investigación Clínicade Cantabria).

2.2. Osteoblast cultures

Bone samples were obtained during hip replacement surgery, in pa-tients with hip fractures (n = 11) or with hip osteoarthritis (n = 9)(mean age 82 ± 5 and 75 ± 7 yr, respectively). Trabecular bone cylin-ders of the central part of the femoral head (thus avoiding the fracturedand the subchondral regions) were obtained with a trephine, cut insmall samples, washed extensively in phosphate-buffered saline andused to set up osteoblast cultures by the primary explant technique(Jonsson et al., 1999). In brief, bone fragments were seeded into T-75plastic flasks containing Dulbecco's modified Eagle's medium, antibi-otics and 10% fetal bovine serum. This allowed osteoblastic precursorcells to migrate from the fragments and proliferate. After confluence,cells were trypsinized and cultured in the appropriate experimentalconditions.

2.3. Immunofluorescence microscopy

Primary osteoblasts obtained from 5 patients with fractures and 4with osteoarthritis were grown on microscope glass coverslips until70–80% confluence. Then they were fixed with 3.7% paraformaldehyde-PBS and permeabilized with 0.5% Triton 100-X for 30 min at roomtemperature. After several washes with PBS and 0.05% PBS-Tween, thecoverslips were incubated overnight at 4 °C with a rabbit polyclonalanti β-catenin antibody (Abcam, Cambridge, UK), diluted 1/200 in PBS.After washing, cells were incubated with a FITC-conjugated anti-rabbitsecondary antibody (Jackson ImmunoResearch Laboratories, WestGrove, PA, USA) for 45 min and mounted with VectaShield (VectorLaboratories, Burlingame, CA, USA). In some experiments, β-cateninimmunolabeling was combined with Texas Red-Phalloidin to stainactin filaments.

Confocalmicroscopywas performedwith a laser scanningmicroscope(LSM 510; Carl Zeiss, Oberkochen, Germany) by using excitation wave-lengths of 488 nm (for FITC) and 543 nm (for Texas Red). Each channelwas recorded independently, and pseudocolor images were generatedand superimposed. TIFF images were transferred to Adobe Photoshop7.0 software (Adobe Systems, San Jose, CA, USA) for presentation.

To measure fluorescence intensities of nuclear and cytoplasmic β-catenin in primary osteoblast cultures from patients with fracturesand osteoarthritis (blindly to the origin of the sample), confocal imagesof at least 130 osteoblasts (without prior selection) of each patientgroup were captured by using a 63× oil 1.4 (NA) objective. Imageswere acquired with no saturated pixels, always using the same confocalsettings, with eightfold averaging at resolution of 1024 × 1024 pixelsand using a pinhole setting of 1. In order to minimize between-day var-iability bias, in each experiment cells of both patient groups wereincluded. Images were background corrected by reference regions out-side the cells and fluorescence intensities were estimated by using theImageJ software (NIH, Bethesda, Maryland, USA; http://rsb.info.nih.gov/ij/). Fluorescence intensities were measured in four regions of in-terest of the same area per nucleus, excluding the nucleolus, and infour regions per cytoplasm. Nuclear and cytoplasmic β-catenin averagevalues and the nucleocytoplasmic ratios were computed for each cellanalyzed and pooled for each patient. The mean values of each patientgroup were then compared by two-tailed unpaired t-test.

2.4. β-Catenin expression

Sub-confluent cultures of first-passage primary osteoblasts wereused to analyze gene expression by reverse transcription (RT) real-


http://rsb.info.nih.gov/ij/

http://rsb.info.nih.gov/ij/


186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

3C. García-Ibarbia et al. / Gene xxx (2013) xxx–xxx

time PCR. The medium was aspirated and fresh medium with 0.1% bo-vine serumalbuminwas added. Forty eight hours later, cellswere rinsedwith phosphate-buffered saline and the RNAwas extracted with Trizol,following manufacturer's instructions (Invitrogen, Carlsbad, CA, USA).The purity and integrity of RNA were checked by absorbance and gelelectrophoresis.

Aliquots of RNAwere reverse-transcribed with the Superscript III kit(Invitrogen), using random hexamers as primers. After RT, the expres-sion of β-catenin was determined by real-time PCR using gene-specific primers and Taqman probes (Applied Biosystems, Foster City,CA, USA) in an ABI7300 apparatus (Applied Biosystems). The amountof PCR product was monitored by fluorescence and the threshold cycle(Ct) for each well was determined. The results were normalized to theexpression of the housekeeping gene TATA box protein (TBP) and thespecific gene expression was calculated as the 2−ΔCt, where ΔCt is thedifference between the gene of interest threshold cycle and the house-keeping threshold cycle.

T

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

EC

2.5. Genetic analysis

The genotype analysis included 353 patients with severe hip osteo-arthritis requiring replacement surgery (182 men and 171 women,age 69 ± 11 yr) and 500 with hip fractures (92 men and 408 women,age 79 ± 12 yr). DNA was isolated from blood or buccal swabs usinga commercial kit (Qiagen, Hilden, Germany) and quantified with Qubittechnique (Invitrogen). Several candidate genes of the Wnt pathwaywere selected on the basis that they showed differential expression inhip osteoarthritis and fractures (Velasco et al., 2010). The gene setincluded 3 Wnt ligands (WNT4, WNT10A and WNT16) and 1 geneencoding a soluble Wnt-binding protein of the frizzled family (SFRP4).We explored the Hapmap database searching SNPs in those geneswith a minimum allelic frequency (MAF) of 10% in the Caucasian popu-lation. Tagging SNPs capturing the most common variants of thesegenes were then selected using the method of Gabriel, implementedin Haploview (Barrett et al., 2005). In addition, we included potentiallyfunctional SNPs identified with the PupaSuite web tool (Conde et al.,2006). The SNP set was genotyped using a Sequenom platform, at theCentro Nacional de Genotipado in Santiago de Compostela, Spain. Repli-cate samples were included for quality control.
R 283
284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

UNCO

R2.6. DNA methylation profiling

Bone samples were obtained from femoral heads removed duringhip replacement as previously reported (Hernandez et al., 2008). Thestudy population included women aged 61–85 years; with osteoarthri-tis (age 73 ± 7 yr; n = 15) or with osteoporotic hip fractures (81 ±3 yr; n = 15). DNAwas isolated from bone samples using phenol:chlo-roform:isoamylalcohol, as previously described (Delgado-Calle et al.,2012c). After bisulphite conversion of DNA using the EZ DNA Methyla-tion kit (Zymo Research, Orange, CA, USA), microarray-based DNAmethylation profiling was performed with the Human MethylationInfinium 27 k DNA Analysis BeadChip (Illumina, San Diego, CA, USA),followingmanufacturer's instructions. This array targets CpG sites locat-ed within the proximal promoter regions of transcription start sites of14,475 consensus coding sequencing (CCDS) in the NCBI Database(Genome Build 36) and 110 miRNA promoters. The assay interrogatesthe loci using two site-specific probes, one designed for the methylatedlocus (M bead type) and another for the unmethylated locus (U beadtype). The methylation level for the interrogated locus is determinedby calculating the ratio of the fluorescent signals from the methylatedvs. unmethylated sites as expressed as beta-values, a quantitative mea-sure of DNA methylation levels of specific CpG that ranges from 0 forcompletely unmethylated to 1 for completely methylated. The analysisincluded 257 CpG sites present in genes related to the Wnt pathway.


ED P

RO

OF

2.7. Gene demethylation and expression

To explore the effects of DNA demethylation on gene expression wetreated the osteoblast-like cell line MG-63 with 5-aza-2-deoxy-azacytidine (AzadC) for 4 days. Three independent cultures (eachincluding control and 1 μM AzadC-treated cells) were used forthese experiments. Gene expression profiling was performed atthe Gene Expression Unit of Genomics Core Facility (SGIKer) of theUniversity of the Basque Country UPV/EHU (Leioa, Spain) usingoligonucleotide-based Agilent Whole Human Genome Oligo Micro-arrays 4×44K G4112F (design ID 014850). Two-color microarray-based gene expression analysis was performed following the QuickAmp Labeling protocol from Agilent Technologies (G4140-90050v5.7; Agilent Technologies España, Las Rozas, Spain). In brief,Cyanine-3 (Cy3) labeled cRNA was prepared from 500 ng of a poolof control RNA samples, and Cyanine-5 (Cy5) labeled cRNA was pre-pared from 300 ng of individual AzadC-treated samples. Dye incor-poration and cRNA yield were monitored with the NanoDrop ND-1000 Spectrophotometer. Aliquots (825 ng) of each Cy3 and Cy5labeled cRNA were fragmented, and co-hybridized to microarraysand washed following manufacturer's recommendations. Slideswere scanned using an Agilent DNA Microarray Scanner G2565BAand the resulting TIFF images were processed with Agilent FeatureExtraction Software v9.5.3. Default parameters for two-color geneexpression microarrays were used for image analysis, data extrac-tion, background correction and dye bias correction. Non-uniformoutliers or signals not significantly above background intensity in70% or more of each channel (Cy3 or Cy5) were filtered out.

2.8. Statistical analyses

The statistical significance of the differences in β-catenin levels be-tween samples from patients with hip fractures and osteoarthritis wastested by the Mann–Whitney tests. Genotype and allele frequencies ofSNPs were tested for consistency with Hardy–Weinberg equilibrium(HWE) proportions using Plink software (Purcell et al., 2007). The allelicfrequency distributions in both groups of patients were comparedassuming additive models with Plink, in the whole population and inthemale and females subgroups. Power analysis of the genetic analyseswas done assuming a log-additive genetic model with Quanto software(available at http://hydra.usc.edu/gxe/).

For methylation, the ratio methylated/unmethylated was estimatedas the beta/(1-beta) ratio and log2-transformed. Data were then nor-malized by the quintile procedure with BRB Array software developedby Dr. Richard Simon and the BRB-ArrayTools Development Team(http://linus.nci.nih.gov/BRB-ArrayTools.html) and between group dif-ferences in the methylation of genes related to the Wnt pathway inKEGG (Kyoto Encyclopedia of Genes and Genomes) were analyzed byt-tests. Significance levels were corrected for multiple testing by themethod of Benjamini to control the FDR (Benjamini and Yekutieli,2005).

To assess the AzadC-induced changes in gene expression,we consid-ered as up-regulated those genes consistently increased in the threeexperiments, with an average fold-change N2; similarly, consistentlydecreased genes with a fold-change b−2 were regarded as down-regulated.

3. Results

3.1. β-Catenin levels and gene expression

In addition to a weak immunostaining throughout the cytoplasm,β-catenin signal was concentrated at the cell cortex, particularly be-neath membrane domains involved in cell–cell interactions. In thislocalization, β-catenin frequently colocalized with the tip of actin fil-aments (Fig. 1). Staining was also found at the nuclei of cells of both


http://hydra.usc.edu/gxe/

http://linus.nci.nih.gov/BRB-ArrayTools.html


TED P

RO

OF

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

1A 1B 1C

2A 2B 2C

Fig. 1. Immunofluorescence of primary osteoblasts grown from fracture (1) or osteoarthritis samples (2). A, β-Catenin staining; B, phalloidin staining of F-actin; C, merged images.


CO

RREC

groups of patients, but the nuclear fluorescence intensity was signifi-cantly lower in cultures grown from patients with hip fractures than inpatients with osteoarthritis (nucleocytoplasmatic ratios of 2.39 ± 0.38versus 3.13 ± 0.38, respectively, p = 0.023). To analyze if β-cateninwas regulated at transcriptional or post-transcriptional levels, we mea-sured its expression in cell cultures. The abundance of β-catenin mRNAwas similar in both groups of patients (Fig. 2).

3.2. Association between gene polymorphisms and disease

The SNPs analyzed and their chromosomal locations are shown inSupplementary Table S1. The genotype distribution did not differ be-tween both groups of individuals (Table 1), neither in the combinedanalysis nor in the sex-stratified analysis. Similar results were obtainedwhen age was included in the analysis as a covariate.

UN

OA Frx40

60

80

100

120 p=0.01

Nu

clea

r β-

cate

nin

Fig. 2. β-Catenin nucleocytoplasmatic staining fluorescence intensity (left panel) and expressiofrom patients with fractures (Frx) or osteoarthritis (OA). Each point represents the result of an


3.3. DNA methylation and gene expression

The analyzed CpG sites and their statistical significance are listed inTable S2 (Supplementary Online Material). Statistical analysis revealed6 CpG sites (out of the 257 related to Wnt pathway genes) with differ-ential methylation in samples from patients with fractures and osteoar-thritis (FDR b 0.1). They included FZD10, CSNK1E, TBL1X, WNT8A, andSFRP4 genes. P-values and the differences in methylation are shown inTable 2. The heatmapof the probemethylation, shown in Fig. 3, revealedimportant heterogeneity among patients with fractures, who tended togroup into two different classes. Within the age range studied, no geneshowed significant age-related differences in methylation.

To explore the potential relationship between DNA methylationand the expression of Wnt-related genes, we treated osteoblast-like MG-63 cells with AzadC, which promotes a global decrease in

OA Frx10-3

10-2

10-1

100

101 p=0.84

β-ca

ten

in e

xpre

ssio

n

n levels assessed by quantitative real-time PCR (right panel) in primary osteoblast culturesindividual patient.



T

PRO

OF

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

Table 1t1:1

t1:2 Allele and genotype frequencies of patients with hip fractures and hip osteoarthritis (homozygotes for theminor allele/heterozygotes/homozygotes for themajor allele), and p-values fort1:3 the differences between both groups of patients.

t1:4 Fracture (n = 500) Osteoarthritis (n = 353)

t1:5 SNP GENE MAF Genotypes MAF Genotypes P (all) P (women) P (men)

t1:6 rs7526484 WNT4 0.23 26/160/284 0.24 18/123/193 0.56 0.46 0.98t1:7 rs2235526 WNT4 0.18 17/130/303 0.18 13/93/217 0.92 0.39 0.65t1:8 rs10917158 WNT4 0.18 17/133/321 0.17 12/90/234 0.70 0.91 0.51t1:9 rs3806557 WNT10A 0.23 30/159/286 0.20 16/108/222 0.18 0.27 0.83t1:10 rs10177996 WNT10A 0.21 27/136/280 0.19 8/103/208 0.19 0.15 0.77t1:11 rs2385199 WNT10A 0.20 24/149/309 0.17 8/104/231 0.14 0.16 0.89t1:12 rs3779381 WNT16 0.26 34/174/258 0.26 22/133/183 0.92 0.88 0.33t1:13 rs2908004 WNT16 0.44 87/252/144 0.42 62/168/115 0.46 0.76 0.94t1:14 rs2707471 WNT16 0.16 13/122/338 0.15 7/87/247 0.65 0.90 0.91t1:15 rs3801385 WNT16 0.07 3/58/412 0.07 4/40/297 0.83 0.65 0.94t1:16 rs2707466 WNT16 0.44 87/241/142 0.42 63/153/117 0.37 0.53 0.92t1:17 rs17143305 WNT16 0.15 12/120/335 0.14 7/82/244 0.58 0.89 0.90t1:18 rs3242 SFRP1 0.38 77/192/183 0.36 46/153/138 0.45 0.88 0.86t1:19 rs1127379 SFRP1 0.44 104/229/159 0.44 64/175/106 0.84 0.66 0.95t1:20 rs7820647 SFRP1 0.39 76/212/173 0.41 53/163/113 0.58 0.31 0.73t1:21 rs11786592 SFRP1 0.35 59/211/206 0.36 44/145/139 0.70 0.54 0.83t1:22 rs6651363 SFRP1 0.40 77/225/177 0.42 61/161/114 0.31 0.58 0.33t1:23 rs10109536 SFRP1 0.41 83/233/176 0.42 66/159/124 0.64 0.42 0.98t1:24 rs17652488 SFRP1 0.51 127/189/121 0.51 94/130/86 0.83 0.62 0.84t1:25 rs10958671 SFRP1 0.13 13/98/377 0.13 6/75/259 0.96 0.50 0.86t1:26 rs17574424 SFRP1 0.19 22/145/321 0.16 6/99/240 0.09 0.15 0.27t1:27 rs7832767 SFRP1 0.05 2/45/425 0.05 1/34/306 0.94 0.63 0.80t1:28 rs968427 SFRP1 0.39 77/222/188 0.38 49/165/134 0.74 0.83 0.27t1:29 rs921142 SFRP1 0.32 53/205/230 0.31 34/142/168 0.57 0.70 0.98

t2:1

t2:2

t2:3

t2:4

t2:5

t2:6

t2:7

t2:8

t2:9

t2:10

t2:11


UNCO

RREC

DNAmethylation. In AzadC-treated cells, 8 genes were up-regulated,whereas 16 were down-regulated (Table 3).

4. Discussion

In this study we found increased nuclear β-catenin in osteoblastsfrom patients with osteoarthritis in comparison with those grownfrom patients with osteoporotic fractures. β-Catenin participates incadherin signaling by binding to the cytoplasmic domain of type Icadherins and linking them to the actin cytoskeleton (Mbalavieleet al., 2006; Nelson and Nusse, 2004). On the other hand, it is a majorplayer in the canonical Wnt pathway. The Wnt pathway regulates thedifferentiation and activity of bone cells, and particularly of the bone-forming cells of the osteoblastic lineage (Williams and Insogna, 2009).On the other hand,Wnt activity has been reported to influence cartilagemetabolism and may be involved in the pathogenesis of osteoarthritis(Corr, 2008; Diarra et al., 2007; Lodewyckx and Lories, 2009; Luytenet al., 2009). Although osteoarthritis has been classically understoodas a cartilage disorder, important bone changes take place in the vicinityof the joints with osteoarthritis. In fact, several lines of evidence suggestthat bonemay playmore than amerely passive role in the pathogenesisof osteoarthritis (Castaneda et al., 2012). Wnt pathway genes are likelyinvolved in both the bone and cartilage alterations that eventually resultin the development of osteoarthritis (Corr, 2008; Kawaguchi, 2009;Lodewyckx and Lories, 2009; Luyten et al., 2009).

We have previously shown a reduced expression of several genes ofthe Wnt pathway in bone tissue samples and osteoblast cultures from

Table 2Wntpathway genes showingdifferentialmethylation of CpG sites.Meanbeta-values (a quantita1 for completely methylated) of samples from patients with osteoarthritis and osteoporotic frtranscription start site (TSS) are shown. FDR, false discovery rate.

Gene CpG location Distance to TSS Osteoarthr

FZD10 47160656 223 0.14CSNK1E 37044362 327 0.29TBL1X 9393597 549 0.19WNT8A 137447943 365 0.54CSNK1A1L 36577573 230 0.64SFRP4 37922543 359 0.15


ED

patients with hip fractures, in comparison with samples from patientswith osteoarthritis (Velasco et al., 2010). Therefore, it could be speculat-ed that the higher Wnt pathway activity may be involved in theincreased bone formation taking place in osteoarthritic joints (causingosteophytes and subchondral bone sclerosis) and perhaps in the differ-ences in bone mass between osteoarthritis and osteoporosis. In thepresent study we confirmed that nuclear β-catenin is more abundantin primary osteoblast cultures grown from osteoarthritis samples thanin samples from patients suffering a hip fracture. This was not accompa-nied by changes in β-catenin gene transcription, which is consistentwith regulation at the posttranscriptional level. Although we did notconfirm increased levels of β-catenin by western-blotting, these resultsare in line with studies using gene expression and reporter vectors thatshowed higher Wnt activity in osteoblasts from patients with osteoar-thritis than in those obtained from fracture cases (Velasco et al., 2010).

Genetic association studies have found Wnt pathway genes, andspecifically Wnt ligands such as WNT16, to be associated with bonemineral density and wrist fractures (Estrada et al., 2012; Medina-Gomez et al., 2012; Zheng et al., 2012). Therefore, we hypothesizedthat the differences inWnt/β-catenin activity could be related to geneticor epigenetic variants. However, we did not find evidence for geneticdifferences between both groups of patients regarding three Wnt li-gands (WNT4, WNT10A and WNT16) or a Wnt inhibitor (SFRP1). Thesenegative results of the genetic association analysis should be interpretedin the context of the limitations inherent to our study. Most important,the aimof our studywas to explorewhether genetic differences contrib-uted to explain the differences inWnt activity between osteoporotic hip

tivemeasure of DNAmethylation levels that ranges from0 for completely unmethylated toactures. The chromosomal location and the distance of the interrogated nucleotide to the

itis (beta) Fractures (beta) P-value FDR

0.08 3.7 × 10−5 0.00940.21 0.00036 0.03850.10 0.00045 0.03850.61 0.00079 0.04280.70 0.00083 0.04280.08 0.00176 0.0755



TED P

RO

OF

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

Osteoarthritis Fractures

Fig. 3. Heat map representation of the methylation of Wnt-pathway (scaled and centered values). Darker color represents higher methylation. Genes with CpG sites showing a trend fordifferential methylation between osteoporosis and osteoarthritis (nominal p-values b0.05) are shown.

t3:1

t3:2

t3:3

t3:4

t3:5

t3:6

t3:7

t3:8

t3:9

t3:10

t3:11

t3:12

t3:13

t3:14

t3:15

t3:16

t3:17

t3:18

t3:19

t3:20


RRECfractures and hip osteoarthritis rather than discovery of genes associat-

ed with these conditions (reason why comparisons with a healthy con-trol group were not done). This was a moderate-sized study, withlimited statistical power. With a type I error of 5%, our study had morethan 80% power to detect disease-associated polymorphisms withodds ratios of 1.4 and 1.6, when the minor allele frequencies are N0.3or N0.1, respectively. However, it was underpowered to detect SNPswith smaller odds ratios. For instance, power to detect alleles with 0.3frequency and odds ratios in the range of 1.1–1.2 would be only 15–41%. In our study sex and age were different in the groups of fracturesand osteoarthritis, which reflects the epidemiological differencesbetween these skeletal disorders. A lack of association between the

UNCO 409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

Table 3Wnt pathway genes up-regulated and down-regulated by AzadC in MG-63 cells(mean ± of three experiments).

Gene Fold-increase Gene Fold-decrease

RAC2 18.1 ± 3.1 SFRP1 7.6 ± 0.9FZD10 12.1 ± 4.4 FZD1 6.2 ± 0.3WNT11 6.2 ± 1.6 WNT5B 5.5 ± 1.2PLCB2 5.0 ± 1.0 CTNNBIP1 4.2 ± 0.7FZD4 4.9 ± 0.6 CTNNB1 4.1 ± 0.8WNT6 3.7 ± 0.1 DKK2 4.0 ± 0.4MYC 3.2 ± 0.1 CSNK2A1 3.2 ± 0.2NLK 2.5 ± 0.1 GSK3B 3.0 ± 0.2

CAMK2G 2.9 ± 0.2PPP2R1A 2.8 ± 0.4PLCB3 2.7 ± 0.3CTBP1 2.7 ± 0.2CACYBP 2.3 ± 0.1RHOA 2.3 ± 0.1CSNK1A1 2.1 ± 0.1FOSL1 2.1 ± 0.1


genotypes and the phenotypes persisted when results were adjustedby age and sex, but the statistical power further decreased underthose analysis conditions. We selected genes on the basis of their differ-ential expression, but we cannot exclude the existence of differences inthe allelic frequency distributions of other genes in theWnt pathway. Infact, some investigators reported an association of osteoarthritis withcertain polymorphisms of the FRZB gene, which encodes secreted friz-zled related protein 3, another Wnt inhibitor (Loughlin et al., 2004).However, this has not been replicated in other reports, including somerecent genome-wide studies (Panoutsopoulou et al., 2011).

Since genetic differences did not explain the differences in Wnt ac-tivity, alternative mechanisms not related to DNA sequence might beinvolved Thus, we hypothesized that epigenetic marks, and specificallycytosine methylation, might underlie the differences in Wnt activitybetween osteoporosis and osteoarthritis. The methylation of cytosinesof CpG dinucleotides is maintained through cell divisions by DNAmethyltransferases. Methylation of CpG-rich sequences of the promoterregions tends to inhibit the transcription of genes known to play impor-tant roles in bone formation and bone resorption. On the other hand, thedemethylation of those CpG-rich regions is associated with the activa-tion of gene expression (Delgado-Calle et al., 2011, 2012b, 2012c). Littleis known about the potential role of CpG methylation in the pathogen-esis of bone changes in osteoarthritis and other skeletal disorders. How-ever, promoter methylation has been demonstrated to modulate Wntpathway activity in other normal and neoplastic tissues (Ekstromet al., 2011; Kocemba et al., 2012). In line with this, in the presentstudywe identified severalWnt-related genes differentiallymethylatedin osteoporosis and osteoarthritis. WNT8A is a Wnt ligand that may bemodulated by estrogen and has been associated with alterations ofbone development, such as cleft palate (Chiquet et al., 2008). Proteinsof the frizzled family, including FZD10 (frizzled family receptor 10),may act as Wnt co-receptors at the cell membranes. On the other



T

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502503504505506507508509510511512513514515516517518519520521522523524525526527528


REC

hand, soluble frizzled-related proteins, including SFRP4 (soluble friz-zled related protein 4), are secreted and may bind Wnt ligands, thuspreventing their interaction with cell membrane receptors (Kawanoand Kypta, 2003; Nakanishi et al., 2007; Wang et al., 2006). TBL1X(transducin (beta)-like 1X-linked) encodes a regulatory protein thatappears to contribute to the regulation of Wnt target genes (Liand Wang, 2008). Casein kinases, including CSNK1E (casein kinase1, epsilon), participate in the regulation of a variety of cell func-tions, and contribute to the signaling cascade initiated by the inter-action of Wnt ligands with their receptors (Valle-Perez et al., 2011).The biological role of CSNK1A1L (casein kinase 1, alpha 1-like) geneis unknown.

In theory, those methylation differences could influence Wnt path-way activity, but further studies are needed to confirm this hypothesis,including detailed analysis of methylation at different nucleotides byother procedures such as pyrosequencing. The methylation signaturesof Wnt-related genes revealed some rather different patterns, suggest-ing that the osteoarthritis and fracture groups may be heterogeneousand include patients with somewhat different pathogenetic mecha-nisms, at least regarding the gene methylation pattern. This is in linewith the views of other investigators that used a clinico-epidemiologicalapproach (Herrero-Beaumont et al., 2009). In general, DNA methylationtends to inhibit gene expression, but this is not a universal phenomenon(Hantusch et al., 2007). In fact, we found that AzadC upregulated someWnt-related genes, but downregulated others. We have previouslyshown that several genes respond similarly to AZadC treatment in prima-ry osteoblasts and in osteoblastic cell lines (Delgado-Calle et al., 2011,2012b). Thus, although in the present study we used an osteoblastic cellline to assess the response to AzadC, the results are likely similar innontransformed osteoblasts. Thus, these experiments support theconcept that DNA methylation-dependent mechanisms influencethe expression of Wnt pathway genes. However, they do not allowestablishing to what extent those changes are the direct conse-quence of the demethylation of the promoters of those genes, orthe result of changes in other regulatory genes upstream in the path-way. For instance, the expression of FZD10 was increased by AzadC,even though its promoter is largely unmethylated in bone samples,with beta-values between 0.08 and 0.14 (see Table 2), and it iseven less methylated in cultured osteoblastic cells (unpublishedresults). Thus, the stimulatory effect of AzadC was likely due to itseffect on another regulatory molecule which in turn stimulatedFZD10 transcription.

529530531532533534535536537538539540541542543544545546547

UNCO

R

5. Conclusion

In conclusion, nuclearβ-catenin levels are higher in osteoblasts formhip osteoarthritis than in osteoblasts from hip fractures. This is in linewith previous reports showing higherWnt pathway activity in osteoar-thritis and may be related to the opposite changes in bone mass andbone formation typical of these disorders. However, the difference inWnt activity is not explained by the allele distribution of common poly-morphisms of various Wnt-related genes. On the other hand, despitesome heterogeneous patterns, several genes in the Wnt pathwaypresented differences inmethylation. Further studies are needed to elu-cidate to what extent those epigenetic differences are involved in thedifferences in Wnt activity.

548549550551552553554555556557558559

Founding source

Supported by grants from Instituto de Salud Carlos III/Fondo deInvestigaciones Sanitarias (FIS 06/0034, 09/0539 and PI12/00615). Thefunding agency has no role in the collection, analysis and interpretationof data; in the writing of the report; or in the decision to submit thearticle for publication.


Conflicts of interest

Authors declare that they do not have conflicts of interest.

Acknowledgment

We acknowledge the excellent technical assistance of CarolinaSañudo andVerónicaMijares.We are grateful to the staff of the Santiagode Compostela Genotyping Center (Centro Español de Genotipado,CEGEN), and particularly to María Torres and Angel Carracedo, fortheir help in thegenotyping study.We also thank technical and scientificsupports provided by SGIker (UPV/EHU, MICINN, GV/EJ, ESF).

OFAppendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.gene.2013.09.080.

ED P

ROReferences

Arcogen Consortium, 2012. Identification of new susceptibility loci for osteoarthritis(arcOGEN): a genome-wide association study. Lancet 380, 815–823.

Arokoski, J.P., Arokoski, M.H., Jurvelin, J.S., Helminen, H.J., Niemitukia, L.H., Kroger, H.,2002. Increased bone mineral content and bone size in the femoral neck of menwith hip osteoarthritis. Ann. Rheum. Dis. 61, 145–150.

Barrett, J.C., Fry, B., Maller, J., Daly, M.J., 2005. Haploview: analysis and visualization of LDand haplotype maps. Bioinformatics 21, 263–265.

Benjamini, Y., Yekutieli, D., 2005. Quantitative trait loci analysis using the false discoveryrate. Genetics 171, 783–790.

Bodine, P.V., et al., 2005. The Wnt antagonist secreted frizzled-related protein-1 controlsosteoblast and osteocyte apoptosis. J. Cell. Biochem. 96, 1212–1230.

Castaneda, S., Roman-Blas, J.A., Largo, R., Herrero-Beaumont, G., 2012. Subchondral boneas a key target for osteoarthritis treatment. Biochem. Pharmacol. 83, 315–323.

Chaganti, R.K., et al., 2010. Bone mineral density and prevalent osteoarthritis of the hip inolder men for the Osteoporotic Fractures in Men (MrOS) Study Group. Osteoporos.Int. 21, 1307–1316.

Chang, J., et al., 2007. Noncanonical Wnt-4 signaling enhances bone regeneration of mes-enchymal stem cells in craniofacial defects through activation of p38 MAPK. J. Biol.Chem. 282, 30938–30948.

Chiquet, B.T., et al., 2008. Variation in WNT genes is associated with non-syndromic cleftlip with or without cleft palate. Hum. Mol. Genet. 17, 2212–2218.

Conde, L., et al., 2006. PupaSuite: finding functional single nucleotide polymorphisms forlarge-scale genotyping purposes. Nucleic Acids Res. 34, W621–W625.

Corr, M., 2008. Wnt-beta-catenin signaling in the pathogenesis of osteoarthritis. Nat. Clin.Pract. Rheumatol. 4, 550–556.

Delgado-Calle, J., Sanudo, C., Sanchez-Verde, L., Garcia-Renedo, R.J., Arozamena, J.,Riancho, J.A., 2011. Epigenetic regulation of alkaline phosphatase in human cells ofthe osteoblastic lineage. Bone 49, 830–838.

Delgado-Calle, J., Garmilla, P., Riancho, J.A., 2012a. Do epigenetic marks govern bonemassand homeostasis? Curr. Genomics 13, 252–263.

Delgado-Calle, J., et al., 2012b. DNAmethylation contributes to the regulation of sclerostinexpression in human osteocytes. J. Bone Miner. Res. 27, 926–937.

Delgado-Calle, J., Sanudo, C., Fernandez, A.F., Garcia-Renedo, R., Fraga, M.F., Riancho, J.A.,2012c. Role of DNA methylation in the regulation of the RANKL-OPG system inhuman bone. Epigenetics 7, 83–91.

Dequeker, J., Aerssens, J., Luyten, F.P., 2003. Osteoarthritis and osteoporosis: clinical andresearch evidence of inverse relationship. Aging Clin. Exp. Res. 15, 426–439.

Diarra, D., et al., 2007. Dickkopf-1 is a master regulator of joint remodeling. Nat. Med. 13,156–163.

Ekstrom, E.J., Sherwood, V., Andersson, T., 2011. Methylation and loss of secreted frizzled-related protein 3 enhances melanoma cell migration and invasion. PLoS One 6,e18674.

Estrada, K., et al., 2012. Genome-wide meta-analysis identifies 56 bone mineral densityloci and reveals 14 loci associated with risk of fracture. Nat. Genet. 44, 491–501.

Fraga, M.F., Esteller, M., 2007. Epigenetics and aging: the targets and the marks. TrendsGenet. 23, 413–418.

Gaur, T., et al., 2005. CanonicalWNT signaling promotes osteogenesis by directly stimulatingRunx2 gene expression. J. Biol. Chem. 280, 33132–33140.

Gordon, M.D., Nusse, R., 2006. Wnt signaling: multiple pathways, multiple receptors, andmultiple transcription factors. J. Biol. Chem. 281, 22429–22433.

Hantusch, B., Kalt, R., Krieger, S., Puri, C., Kerjaschki, D., 2007. Sp1/Sp3 and DNA-methylation contribute to basal transcriptional activation of human podoplanin inMG63 versus Saos-2 osteoblastic cells. BMC Mol. Biol. 8, 20.

Hernandez, J.L., et al., 2008. Aromatase expression in osteoarthritic and osteoporotic bone.Arthritis Rheum. 58, 1696–1700.

Herrero-Beaumont, G., Roman-Blas, J.A., Castaneda, S., Jimenez, S.A., 2009. Primary osteo-arthritis no longer primary: three subsets with distinct etiological, clinical, and ther-apeutic characteristics. Semin. Arthritis Rheum. 39, 71–80.





560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589Q4590591592593

594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627

629


Jonsson, K.B., Frost, A., Nilsson, O., Ljunghall, S., Ljunggren, O., 1999. Three isolation tech-niques for primary culture of human osteoblast-like cells: a comparison. Acta Orthop.Scand. 70, 365–373.

Kawaguchi, H., 2009. Regulation of osteoarthritis development by Wnt-beta-catenin sig-naling through the endochondral ossification process. J. Bone Miner. Res. 24, 8–11.

Kawano, Y., Kypta, R., 2003. Secreted antagonists of theWnt signalling pathway. J. Cell Sci.116, 2627–2634.

Kocemba, K.A., et al., 2012. Transcriptional silencing of the Wnt-antagonist DKK1 by pro-moter methylation is associated with enhanced Wnt signaling in advanced multiplemyeloma. PLoS One 7, e30359.

Li, J., Wang, C.Y., 2008. TBL1-TBLR1 and beta-catenin recruit each other to Wnt target-gene promoter for transcription activation and oncogenesis. Nat. Cell Biol. 10,160–169.

Lodewyckx, L., Lories, R.J., 2009.WNT Signaling in osteoarthritis and osteoporosis: what isthe biological significance for the clinician? Curr. Rheumatol. Rep. 11, 23–30.

Loughlin, J., et al., 2004. Functional variants within the secreted frizzled-related protein 3gene are associated with hip osteoarthritis in females. Proc. Natl. Acad. Sci. U. S. A.101, 9757–9762.

Luyten, F.P., Tylzanowski, P., Lories, R.J., 2009. Wnt signaling and osteoarthritis. Bone 44,522–527.

MacDonald, B.T., Tamai, K., He, X., 2009. Wnt/beta-catenin signaling: components, mech-anisms, and diseases. Dev. Cell 17, 9–26.

Mbalaviele, G., Shin, C.S., Civitelli, R., 2006. Cell–cell adhesion and signaling throughcadherins: connecting bone cells in their microenvironment. J. Bone Miner. Res. 21,1821–1827.

Medina-Gomez, C., et al., 2012. Meta-analysis of genome-wide scans for total body BMDin children and adults reveals allelic heterogeneity and age-specific effects at theWNT16 locus. PLoS Genet. 8, e1002718.

Nakanishi, R., et al., 2007. Osteoblast-targeted expression of Sfrp4 in mice results in lowbone mass. J. Bone Miner. Res.

Nelson, W.J., Nusse, R., 2004. Convergence of Wnt, β-catenin, and cadherin pathways.Science 303, 1483–1487.

Panoutsopoulou, K., et al., 2011. Insights into the genetic architecture of osteoarthritisfrom stage 1 of the arcOGEN study. Ann. Rheum. Dis. 70, 864–867.

UNCO

RRECT

628


PRO

OF

Piters, E., Boudin, E., Van Hul, W., 2008. Wnt signaling: a win for bone. Arch. Biochem.Biophys. 473, 112–116.

Purcell, S., et al., 2007. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. 81, 559–575.

Ralston, S.H., 2010. Osteoporosis as an hereditary disease. Clin. Rev. BoneMiner. Metab. 8,68–76.

Riancho, J.A., et al., 2011. Wnt receptors, bone mass, and fractures: gene-wide associationanalysis of LRP5 and LRP6 polymorphisms with replication. Eur. J. Endocrinol. 164,123–131.

Richards, J.B., et al., 2009. Collaborative meta-analysis: associations of 150 candidategenes with osteoporosis and osteoporotic fracture. Ann. Intern. Med. 151, 528–537.

Styrkarsdottir, U., et al., 2009. New sequence variants associated with bone mineral density.Nat. Genet. 41, 15–17.

Valdes, A.M., et al., 2007. Sex and ethnic differences in the association of ASPN, CALM1,COL2A1, COMP, and FRZB with genetic susceptibility to osteoarthritis of the knee.Arthritis Rheum. 56, 137–146.

Valero, C., et al., 2011. Relationship of sclerostin and secreted frizzled protein polymor-phisms with bonemineral density: an association studywith replication in postmen-opausal women. Menopause 18, 802–807.

Valle-Perez, B., Arques, O., Vinyoles, M., de Herreros, A.G., Dunach, M., 2011. Coordinatedaction of CK1 isoforms in canonical Wnt signaling. Mol. Cell. Biol. 31, 2877–2888.

van Amerongen, R., Nusse, R., 2009. Towards an integrated view of Wnt signaling indevelopment. Development 136, 3205–3214.

Velasco, J., et al., 2010. Wnt pathway genes in osteoporosis and osteoarthritis: differentialexpression and genetic association study. Osteoporos. Int. 21, 109–118.

Wang, H.Y., Liu, T., Malbon, C.C., 2006. Structure–function analysis of Frizzleds. Cell. Signal.18, 934–941.

Williams, B.O., Insogna, K.L., 2009. Where Wnts went: the exploding field of Lrp5 andLrp6 signaling in bone. J. Bone Miner. Res. 24, 171–178.

Yao, W., Cheng, Z., Shahnazari, M., Dai, W., Johnson, M.L., Lane, N.E., 2010. Overexpressionof secreted frizzled-related protein 1 inhibits bone formation and attenuates parathy-roid hormone bone anabolic effects. J. Bone Miner. Res. 25, 190–199.

Zheng, H.F., et al., 2012. WNT16 influences bone mineral density, cortical bone thickness,bone strength, and osteoporotic fracture risk. PLoS Genet. 8, e1002745.

ED



Documents

Contribution of genetic and epigenetic mechanisms to Wnt